I think you may be misunderstanding what the man is saying because he’s couching it in the kinds of terms you usually see in courtrooms. What he’s saying is that in an audio of this quality you would expect it to match up about 90% with a voice exemplar (sample of the person’s voice) to be considered a forensic match. Since it doesn’t match 90% he’s saying that with what he considers to be a reasonable degree of scientific certainty, it is NOT Zimmerman
You can argue with his conclusion, but that’s what he’s saying
he says in the OS article he can say with scientific certainty it is not Zim because it does not reach 90% (his own site says comparing these two samples is useless by the way)
then he told CNN that 60% would justify "confidence" that it is Zimmerman.
The guy is an idiot.